PXE boot and NFS problem in Plop 4.2.2

Started by The_Raven, June 16, 2012, 20:46:09 PM

Previous topic - Next topic

The_Raven

Hello everybody,
i have a problem with plop 4.2.2 and PXE boot. Before i use plop 4.2.1 and everything works ok. I think it's a bug, but not sure.
If i boot from PXE, the kernel and initrd loads and startup. Then plop trys to connect to NFS share and this does not work. I get a lot of errors like this:
svc: failed to register lockdv1 RPC service errno 111
If i take smbfs instead of NFS it works ok. And the strange thing is, if i boot with smbfs and then try to mount the nfs, it works.
I search the whole i-net about this error. Most people say that it has something to do with ipv6, but not sure if this is true or not. But can not beleave this, because if plop is running (with smbfs or from usb) mounting of nfs works. Only at boot it seems not to work.

THX for help and greets

Raven

Elmar

hi,
what is happening when you use for plop linux 4.2.2 the initrd of 4.2.1?

best regards
elmar

The_Raven

#2
Hm, sorry for posting in wrong forum.
Ok, good idea with initrd. But now i realise that something is wrong, because 4.2.1 makes the same error!  :o
Tested 4.2.2, 4.2.1 and 4.2.0 always the same problem (with kernel and initrd).
But maybe i mean 4.1.2 and not 4.2.1. But can not find 4.1.2 on the plop server. Where can i download it? I like to test if it works with 4.1.2.
On the server side i see this message:
mountd[1813]: authenticated mount request from 10.0.0.100:897 for /srv/tftp/ploplinux (/srv/tftp/ploplinux)

Edit:
Ok, found 4.1.2 and tested:
Kernel and initrd from 4.1.2 ->  mount ok but plop not found (think this is ok, because data is from 4.2.2)
Testet with 4.2.2 kernel and initrd 4.1.2 -> mount ok but plop not found (think this is ok, because data is from 4.2.2)
Kernel 4.1.2 and initrd 4.2.2 -> system hangs before nfs share is mounted. (kernel panic, FATAL: Kernel to old)
So it works on 4.1.2 but not in 4.2.0/1/2. It looks like the failure is in initrd. But strange is that no one else has this problem.  :-\

Elmar

hmm, which linux do you use as server? do you know what nfs version is installed?

The_Raven

The server is a debian squeeze 6.0.4 with kernel 2.6.32-5-686. And the nfs-server is:
nfs-kernel-server 1:1.2.2-4squeeze2

And the share loks like this:
/srv/tftp/ploplinux/ 10.0.0.0/24(ro,sync,no_subtree_check)

Elmar

when you mount from the command line, do you use any additional parameters?

for a test boot plop linux with the parameter "busybox" and try to mount the nfs share.

The_Raven

Ok, booted plop with smbfs and try to mount nfs like this:
mount 10.0.0.1:/srv/tftp/ploplinux /mnt
Works perfect. No error and i can browse in to the directory and see all the stuff.

Then booted plop with kernel parameter "busybox".
And mount the nfs the same way as before. Works.

Elmar

i found and fixed a problem with the busybox parameter. download http://download.plop.at/tmp/initramfs.gz

when the busybox shell has been started, then you see only the "#"
try the nfs mount with "mount" and "mount.nfs"

The_Raven

Ok, now busybox works.
After setting up the network:
try to mount with mount -> no such file or directory
with mount.nfs -> svc: failed to register lockdv1 RPC service errno 111

I have do leave now, but test it this evening more. THX for the help!  :)

Elmar

ok, i think i know what the problem is

see you later

Elmar

it should be fixed now. use (same link) http://download.plop.at/tmp/initramfs.gz

when i tested 4.2.x, the nfs boot worked fine, but the nfs server had an old nfs version. meanwhile i updated all servers and it stopped working. now i fixed it with starting portmap before the nfs mount and it works fine. the lockdv1 warning had nothing to do with the mount problem. you can ignore it.

the fix will be in the upcoming plop linux version

best regards
elmar

The_Raven

Ok, tested with the new initrd and... It works well! :)
The error message is still here, but as you sayed before its just a warning.

So thank you for fixing this so fast.  :)

Elmar